Geek Answers Handbook

Wrapper Function Selection Methods (Machine Learning) At Scikit Learn

I am trying to decide between learning scikit and a Weka data mining tool for my machine learning project. However, I realized the need to select features. I would like to know if scikit learn has wrapper methods for function selection.

+5
source share
2 answers

scikit-learn supports Recursive Functions Removal (RFE), which is a wrapper method for selecting functions.

mlxtend , a stand-alone Python library that works well with scikit-learn, also provides a Selective Function Selector (SFS) that works differently:

RFE is less computationally complex using weighting coefficients (e.g. linear models) or feature characteristics (tree-based algorithms) for recursive replication of functions, while SFS excludes (or adds) functions based on a custom classifier / regression performance metric .

+5
source

You can find examples for:

  • Step forward
  • Step back Function selection
  • Comprehensive selection of features

Here

0
source

More articles:


All Articles